Zurich Open Repository and Archive University of Zurich Main Library Strickhofstrasse 39 CH-8057 Zurich www.zora.uzh.ch

Year: 2012

Histopathological patterns of : a phosphate type can be distinguished from a type

Wiech, T ; Hopfer, H ; Gaspert, A ; Banyai-Falger, S ; Hausberg, M ; Schröder, J ; Werner, M ; Mihatsch, M J

Abstract: BACKGROUND: The etiology of nephrocalcinosis is variable. In this study, we wanted to elucidate whether the histopathological appearance of calcium phosphate deposits provides information about possible etiology. METHODS: Autopsy cases from the years 1988 to 2007 and native kidney from a 50-year period (1959-2008) with nephrocalcinosis were identified. The cases were re-evaluated by light microscopy. The autopsy cases were analysed according to the underlying disease. The biopsy cases were grouped with respect to the likely etiology of nephrocalcinosis. Total number, den- sity, localization, size and pattern of all foci were documented and correlated with clinical and laboratory data. RESULTS: About 223 of 12 960 autopsy cases (1.7%) had nephrocalcinosis, 111 of which (49.8%) suffered from advanced malignant tumours. Nephrocalcinosis was the main diagnosis in 48 of 12 480 native kidney biopsies (0.4%). Clinicopathological correlation revealed a specific pat- tern of calcification associated with hyperphosphataemia and/or hyperphosphaturia: these cases showed predominant globular or shell-like (phosphate type). In contrast, the biopsies of thehy- percalcaemic/hypercalciuric group had a different predominant pattern with clumpy or finely granular calcifications (calcium type). CONCLUSIONS: Our results indicate that hyperphosphaturia-associated cases of nephrocalcinosis can be distinguished from hypercalciuria-associated cases histopathologically.

DOI: https://doi.org/10.1093/ndt/gfr414

Posted at the Zurich Open Repository and Archive, University of Zurich ZORA URL: https://doi.org/10.5167/uzh-50420 Journal Article Published Version

Originally published at: Wiech, T; Hopfer, H; Gaspert, A; Banyai-Falger, S; Hausberg, M; Schröder, J; Werner, M; Mihatsch, M J (2012). Histopathological patterns of nephrocalcinosis: a phosphate type can be distinguished from a calcium type. , Dialysis, Transplantation, 27(3):1122-1131. DOI: https://doi.org/10.1093/ndt/gfr414 NDT Advance Access published July 29, 2011

Nephrol Dial Transplant (2011) 0: 1–10 doi: 10.1093/ndt/gfr414

Original Article

Histopathological patterns of nephrocalcinosis: a phosphate type can be distinguished from a calcium type

Thorsten Wiech1,2, Helmut Hopfer3, Ariana Gaspert4, Susanne Banyai-Falger5, Martin Hausberg6, Josef Schro¨der7, Martin Werner2 and Michael J. Mihatsch3

1Centre of Chronic Immunodeficiency, University Medical Center Freiburg, University of Freiburg, Freiburg, Germany, 2Institute of Pathology, University Hospital Freiburg, Freiburg, Germany, 3Institute for Pathology, University Hospital Basel, Basel, Switzerland, 4 5 6

Institute of Surgical Pathology, University Hospital Zurich, Zurich, Switzerland, Klinik St Anna, Luzern, Switzerland, Sta¨dtisches Downloaded from Klinikum Karlsruhe, Karlsruhe, Germany and 7Institute of Pathology, University Hospital Regensburg, Regensburg, Germany Correspondence and offprint requests to: Thorsten Wiech; E-mail: [email protected]

Abstract associated with nephrolithiasis. Pathologists distinguish be- Background. The etiology of nephrocalcinosis is variable. tween oxalosis and nephrocalcinosis, the latter again cover- http://ndt.oxfordjournals.org/ In this study, we wanted to elucidate whether the histopa- ing different compositions but excluding calcium oxalate. thological appearance of calcium phosphate deposits Traditionally, pathologists used the term ‘calcium nephro- provides information about possible etiology. sis’ for dystrophic calcification i.e. calcification of necrosis Methods. Autopsy cases from the years 1988 to 2007 and and the term ‘nephrocalcinosis’ for metastatic calcification native kidney biopsies from a 50-year period (1959–2008) [1]. In our study, we defined nephrocalcinosis as metastatic with nephrocalcinosis were identified. The biopsy cases renal calcifications (excluding oxalosis) by histopatholog- were re-evaluated by light microscopy. The autopsy cases ical detection of calcium deposits without associated tissue were analysed according to the underlying disease. The bi- necrosis. at Universitaetsspital on October 27, 2011 opsy cases were grouped with respect to the likely etiology Nephrocalcinosis can be classified according to the local- of nephrocalcinosis. Total number, density, localization, ization of the deposits: calcium can precipitate predomi- size and pattern of all calcification foci were documented nantly in the medulla, as for example in the medullary and correlated with clinical and laboratory data. sponge kidney. Calcifications mainly located in the renal Results. About 223 of 12 960 autopsy cases (1.7%) had cortex are found, for example, in cortical necrosis. In con- nephrocalcinosis, 111 of which (49.8%) suffered from ad- trast to ultrasonography or computed tomography, light mi- vanced malignant tumours. Nephrocalcinosis was the main croscopy can further distinguish whether calcifications are diagnosis in 48 of 12 480 native kidney biopsies (0.4%). localized in the tubular lumens, tubular epithelium, tubular Clinicopathological correlation revealed a specific pattern basement membrane or in interstitial space [2]. On the other of calcification associated with hyperphosphataemia and/or hand, nephrocalcinosis can be classified according to the hyperphosphaturia: these cases showed predominant glob- chemical composition, which is likely associated with the ular or shell-like calcifications (phosphate type). In con- etiology of renal calcification. Typically, calcium precipi- trast, the biopsies of the hypercalcaemic/hypercalciuric tates with anions in the kidney forming calcium oxalate or group had a different predominant pattern with clumpy or finely granular calcifications (calcium type). calcium phosphate (as hydroxyl apatite) deposits. Conclusions. Our results indicate that hyperphosphaturia- Renal calcium deposits occur (i) in disorders associated associated cases of nephrocalcinosis can be distinguished with impaired homeostasis and an increased renal excretion from hypercalciuria-associated cases histopathologically. of calcium or phosphate or (ii) by an altered solubility, e.g. due to a locally altered pH value in the course of renal Keywords: ; histopathology; hypercalciuria; tubular acidosis, hypocitraturia or tissue necrosis. Apart hyperphosphaturia; nephrocalcinosis from that, factors facilitating crystal adhesion to the tubular epithelium seem to be a prerequisite for nephrocalcinosis [3–5]. Hypercalciuria can occur with hypercalcaemia (see Introduction Table 1, e.g. in primary hyperparathyroidism, elevated vi- tamin D3, sarcoidosis) or with normal serum calcium levels Nephrocalcinosis is a clinical term defined as the renal (idiopathic hypercalciuria). In analogy, hyperphosphaturia deposition of calcium. This definition includes deposits can be associated with hyperphosphataemia or with normal of different chemical composition, such as calcium oxalate or low serum phosphate levels due to impaired reabsorption or calcium phosphate, and can, but does not have to be in the tubules.

Ó The Author 2011. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. For Permissions, please e-mail: [email protected] 2 T. Wiech et al. Table 1. Clinicopathological data of 48 biopsy cases with nephrocalcinosisa

Gender Age Calcification Calcification Clumpy Granular Globular Shell- Onion (M/F) (years) Disease/condition type (Ca/P) foci (number) (%) (%) (%) like (%) skin (%)

Hypercalciuria M 52 Lung cancer Ca 10 10.0 60.0 10.0 20.0 0.0 M 3 Lymphoblastic Ca 5 40.0 60.0 0.0 0.0 0.0 leukemia M 25 Sarcoidosis Ca 15 13.3 66.7 20.0 0.0 0.0 M 51 Sarcoidosis Ca 19 15.8 36.8 31.6 10.5 5.3 M 49 Sarcoidosis Ca 5 40.0 60.0 0.0 0.0 0.0 F 65 Hyperparathyreoidism Ca 6 16.7 83.3 0.0 0.0 0.0 F 34 Hyperparathyreoidism Ca 12 25.0 66.7 8.3 0.0 0.0 M 43 Paracellin 1 mutation Ca 5 60.0 40.0 0.0 0.0 0.0 M 13 Dent’s disease Ca 4 0.0 100.0 0.0 0.0 0.0 Hyperphosphaturia F 62 Phosphate purgatives P 11 0.0 0.0 9.1 72.7 18.2 M 75 Phosphate purgatives P 27 3.7 14.8 29.6 40.7 11.1 F 67 Phosphate purgatives P 3 0.0 0.0 66.7 33.3 0.0 F 69 Phosphate purgatives P 29 0.0 0.0 37.9 62.1 0.0 F 78 Phosphate purgatives P 18 11.1 11.1 27.8 50.0 0.0

F 49 Phosphate P 9 22.2 11.1 44.4 22.2 0.0 Downloaded from M 10 Ca/P 6 0.0 50.0 50.0 0.0 0.0 F 2 Nephrotic syndrome P 23 0.0 26.1 34.8 39.1 0.0 M 1 Nephrotic syndrome P 8 12.5 25.0 12.5 50.0 0.0 Unknown cause F 69 Unknown Ca 13 0.0 53.8 7.7 23.1 15.4 M 23 Unknown P 9 33.3 0.0 22.2 44.4 0.0 M 46 Unknown Ca 1 100.0 0.0 0.0 0.0 0.0 F 42 Unknown P 3 33.3 0.0 66.7 0.0 0.0 http://ndt.oxfordjournals.org/ M Unknown Ca 149 73.8 26.2 0.0 0.0 0.0 F 2 Unknown Ca 36 61.1 33.3 0.0 2.8 2.8 M 18 Unknown Ca 5 80.0 20.0 0.0 0.0 0.0 M 37 Unknown P 15 13.3 6.7 13.3 66.7 0.0 F 44 Unknown P 11 18.2 27.3 18.2 36.4 0.0 F 69 Unknown P 29 6.9 0.0 17.2 75.9 0.0 F 39 Unknown P 9 11.1 22.2 0.0 66.7 0.0 M 29 Unknown Ca 22 63.6 27.3 4.5 0.0 4.5 F 50 Unknown Ca 3 33.3 33.3 33.3 0.0 0.0

F 30 Unknown P 9 0.0 44.4 22.2 33.3 0.0 at Universitaetsspital on October 27, 2011 F 51 Unknown Ca/P 4 50.0 0.0 0.0 50.0 0.0 F 61 Unknown P 21 0.0 23.8 47.6 28.6 0.0 F 69 Unknown Ca 10 20.0 50.0 0.0 30.0 0.0 F 53 Unknown P 9 22.2 11.1 66.7 0.0 0.0 F Unknown P 8 0.0 25.0 0.0 75.0 0.0 M Unknown P 7 28.6 14.3 0.0 57.1 0.0 M 40 Unknown Ca 3 100.0 0.0 0.0 0.0 0.0 M 9 Unknown P 33 9.1 33.3 24.2 33.3 0.0 F 32 Unknown Ca/P 4 0.0 50.0 50.0 0.0 0.0 M 62 Unknown Ca 5 60.0 0.0 40.0 0.0 0.0 M 42 Unknown Ca 8 37.5 25.0 0.0 37.5 0.0 F 66 Unknown Ca 8 62.5 0.0 0.0 37.5 0.0 M 78 Unknown Ca 7 71.4 0.0 28.6 0.0 0.0 F 69 Unknown Ca 4 25.0 75.0 0.0 0.0 0.0 M 36 Unknown Ca 41 12.2 46.3 17.1 24.4 0.0 M 73 Unknown P 11 9.1 0.0 63.6 27.3 0.0 aOverview of the findings in all 48 biopsy cases: column 4 shows the classification of the respective cases into ‘Ca’ (calcium type), if the percentage of clumpy plus finely granular calcifications is >50% and ‘P’ (phosphate type) and if the percentage of globular plus shell-like calcifications is >50%. Three cases could not be classified (Ca/P) because they showed 50% calcium type and 50% phosphate type patterns. Column 5 shows the total number of calcification foci found in the biopsy. Columns 6–10 show the respective frequency of the five different calcification patterns.

Elevated levels of calcium as well as elevated levels of dystrophic calcification of necrosis) in autopsies and re- phosphate can lead to calcium phosphate deposits, which nal biopsies and (ii) to elucidate whether the etiology of can be visualized histochemically by the von Kossa stain. nephrocalcinosis was associated with a special morpho- Hyperphosphataemia can be caused by oral sodium phos- logical pattern, including appearance, distribution and phate bowel purgative intake, leading to renal calcium density of calcifications. A correlation would indicate that phosphate deposition, also referred to as acute phosphate the histological examination of calcification patterns nephropathy [6–8]. Hyperphosphataemia has also been re- could help to determine the possible mechanism and ported in children suffering from nephrotic syndrome [9]. etiology of nephrocalcinosis in a given case, which is The aim of this study was (i) to determine the fre- importantinbiopsycasesif nephrocalcinosis is an quency of nephrocalcinosis (excluding oxalosis and incidental finding. Histopathological patterns of nephrocalcinosis 3 Materials and methods of the cases showed the calcium type and the other half showed the phosphate type of calcification (see below). Patients and specimens Frequency. The frequency of nephrocalcinosis was determined in autopsy cases of the years 1988–2007 (n ¼ 12 960) and in biopsies of Morphology native kidneys of the years 1959–2008 (n ¼ 12 480). Five different patterns of calcifications could be distin- Morphology. All biopsies with the leading histopathological diagnosis guished: clumpy, finely granular, globular, shell-like and nephrocalcinosis of the years 1959–2008 were included in this study. All onion skin-like (see Figure 2). The overall frequencies (of available information, such as age, sex, clinical history, symptoms and laboratory parameters was documented. Histological slides of the unde- total 742 calcifications) of the different patterns were calcified formalin (4%) fixed, paraffin-embedded biopsies were systemati- 30.1% clumpy, 30.1% finely granular, 15.8% globular, cally analysed in haematoxylin and eosin (H&E) and periodic acid Schiff 22.8% shell-like and 1.3% onion skin-like. H&E stain (PAS) stains as well as von Kossa and Alizarin stains for visualization of proved to be the best for the differentiation of the calcifi- calcium phosphate depositions. Immunohistochemical staining of Tamm- Horsfall protein (Anti-human Tamm-Horsfall Protein monoclonal anti- cation patterns. Neither von Kossa stain nor Alizarin stain body clone 10.32; Cedarlane laboratories Ltd, Ontario, Canada. Dilution allowed the differentiation of the different patterns of 1:16000) was performed on paraffin sections of selected cases with calcification. sufficient shell-like calcification. Median age of the patients (n ¼ 48) was 44.9 years Quantitative and semiquantitative evaluation (1–74.8 years) and half of the patients were females

Firstly, the biopsy area on the slide (square millimeters) was measured and (24 of 48). Mean serum was 218 (19–800) Downloaded from the proportion of renal cortex and medulla (%) was assessed in order to lmol/L, mean serum calcium 2.55 (2.0–4.1) mmol/L and calculate the respective density of calcifications. The density was defined mean serum phosphate was 1.31 (0.9–3.1) mmol/L. as the number of calcification foci per square millimeters of the biopsy and The biopsies contained 0 to 97 (median 15) glomeruli. calculated by dividing the total number of calcification foci found in the biopsy by the total area of the biopsy measured on the slide (foci per square The percentage of globally obsolescent glomeruli ranged millimeters). The total number of glomeruli and number of obsolescent from 0 to 75% (median 11%). Tubular atrophy and inter- glomeruli were determined and the percentage of the total area with tubular stitial fibrosis varied from 0 to 95% (median 10%). The http://ndt.oxfordjournals.org/ atrophy and interstitial fibrosis was estimated. The total number of calci- density of calcification foci ranged from 0.01 to 14.9 foci fication foci were counted and documented with respect to the following per square millimeters (average 1.65/mm2), 0 to 36.7 foci/ features: localization (cortex, medulla), compartment (tubular lumen, tub- 2 2 ular epithelium, tubular basement membrane, interstitium or glomerulus), mm in the cortex and 0 to 5.6 foci/mm in the medulla. In size (smaller than a tubular cell, as large as a single cell, larger than a cell, 32 biopsies containing renal cortex and medulla, it was larger than a tubular cross section) and structure. The morphological pat- possible to evaluate the predominant location of nephrocal- terns of calcification were analysed in detail. The results were expressed as cinosis. In 18 of these 32 cases (56%), the calcification percent of all calcification foci in the individual biopsy. density was higher in the medulla and in 14 cases (44%), Energy-filtered electron microscopy

the density was higher in the cortex. The number of calci- at Universitaetsspital on October 27, 2011 In order to further characterize the composition of the shell-like calcifica- fication foci adjusted according to the mean biopsy area tions observed in the phosphate type of nephrocalcinosis, energy-filtered (15.4/mm2) was 25.4 foci on average. electron microscopy was performed. This technique, described in detail Total calcification density did not correlate significantly elsewhere [10], provides high-resolution imaging with specific element ¼ detection and localization [11]. Shell-like calcifications were examined with serum creatinine levels [correlation coefficient r using this method in order to determine the chemical composition, partic- 0.37 (P ¼ 0.1904, 95% CI 0.20–0.75)] nor with calcium ularly the presence of calcium, phosphor and oxygen within the optically 3phosphate product [r ¼ 0.27 (P ¼ 0.28, 95% CI 0.23– empty centres. 0.65)]. No significant correlation was found between serum Statistical analysis calcium and serum creatinine [r ¼ 0.18 (P ¼ 0.3987, 95% Statistical analysis was done with Prism 5 (GraphPad Software, Inc, La CI 0.24–0.54)]. However, there was a significant Jolla, CA) using the non-parametric Mann–Whitney test. The morpholog- correlation between serum phosphate and serum creatinine ical and clinical parameters were correlated with each other using [r ¼ 0.84 (P ¼ 0.0001, 95% CI 0.57–0.94)] and between Pearson’s correlation coefficient. the calcium 3phosphate product and serum creatinine [r ¼ 0.88 (P < 0.0001, 95% CI 0.67–0.96)]. Results The likely etiology of nephrocalcinosis in our biopsy cases could be identified in 18 of 48 cases: malignant tu- mours (n ¼ 2; one acute lymphoblastic leukemia and one Frequency metastatic lung cancer), sarcoidosis (n ¼ 3), primary hyper- In 223 of 12 960 (1.7%) autopsy cases from the years 1988 parathyroidism (n ¼ 2), paracellin 1 mutation (n ¼ 1), to 2007, nephrocalcinosis was a noteworthy finding. Dent’s disease (n ¼ 1), intake of sodium phosphate bowel Around 111 of the patients (49.8%) had advanced malig- purgatives before colonoscopy (n ¼ 5), children with neph- nant tumours (Figure 1A), in only two of the others, the rotic syndrome (n ¼ 3) and phosphate diabetes (n ¼ 1). In etiology could be determined; one patient suffered from the remaining cases, the etiology remained unclear even sarcoidosis and one from primary hyperparathyroidism. after careful re-evaluation of the clinical records (Table 1). In 48 of 12 480 native cases (0.4%), neph- These 18 cases with known etiology were then divided rocalcinosis was a leading diagnosis (Table 1). In the last into two major groups according to the likely mechanism: of 10 five-year periods (2004–2008), there was an in- (i) diseases and conditions associated with hypercalcaemia crease in the number of biopsies with nephrocalcinosis and/or hypercalciuria (malignant tumours, sarcoidosis, (19 cases), compared to the earlier 5-year periods (0–8 hyperparathyroidism, paracellin 1 mutation, Dent’s dis- cases, Figure 1B). As illustrated in Figure 1B, about half ease) defined as the ‘calcium group’ and (ii) diseases and 4 T. Wiech et al. Downloaded from http://ndt.oxfordjournals.org/ at Universitaetsspital on October 27, 2011

Fig. 1. Frequency of nephrocalcinosis. (A) Autopsy cases: 111 of 223 (49.7%) cases of nephrocalcinosis were associated with malignant tumours. Twenty-five of them had haematological neoplasia, followed by lung (19 cases) and breast (17 cases) cancer, most of which had bone metastases. (B) Frequency of nephrocalcinosis in native kidney biopsies over the years. Note: there was a clear increase in the last 5 years (19 cases in 2004–2008). About half of the cases showed a calcium type and the other half a phosphate type calcification pattern. In three cases, the calcification type could not be assigned to the calcium or phosphate type since half of the calcifications within the respective biopsies showed calcium type patterns and half showed phosphate type patterns. conditions associated with hyperphosphataemia and/or hy- the clinical records. Thus, since hyperphosphaturia can oc- perphosphaturia (sodium phosphate bowel purgatives, cur in children with nephrotic syndrome [9], we classified nephrotic syndrome in children and phosphate diabetes) these cases as members of the phosphate group. In regard to defined as the ‘phosphate group’. All three children with the predominant morphological patterns of the calcium and nephrotic syndrome had normal serum calcium levels and phosphate group, a striking difference could be observed. no other reason for nephrocalcinosis could be identified in In the calcium group, the clumpy and the finely granular Histopathological patterns of nephrocalcinosis 5 Downloaded from http://ndt.oxfordjournals.org/ at Universitaetsspital on October 27, 2011

Fig. 2. Histopathological patterns of nephrocalcinosis. (A) Five different histopathological patterns (H&E 3400, scale bars 20 lm) and (B) schematic illustration: Column 1, clumpy pattern with homogenous calcifications; column 2, finely granular pattern; column 3, globular pattern with small hollow spherules; column 4, shell-like pattern with thin outer calcifications surrounding empty cores and column 5, onion skin-like pattern with multilayered calcifications. (C) The group of ‘hypercalciuria’ showed predominant calcification patterns 1 and 2 (clumpy and finely granular). In contrast, cases of ‘hyperphosphaturia’ had a predominance of patterns 3 and 4 (globular and shell-like). Only very few onion skin-like calcifications were found. (D) After combining patterns 1 1 2 and 3 1 4, significant differences could be observed: patterns 1 1 2 were associated with conditions leading to hypercalciuria (calcium type) (P ¼ 0.0003), whereas patterns 3 1 4 were associated with conditions leading to hyperphosphaturia (phosphate type) (P ¼ 0.0004). patterns predominated. In contrast, in the phosphate group, nephrocalcinosis. These grouped patterns again showed the most frequent pattern was the shell-like calcification, significant differences between both major groups using followed by the globular pattern. Patterns 1–4 reached stat- the Mann–Whitney test (Figure 2). istical significance, when the frequencies in the two major In order to assess how specific the phosphate type of disease groups were compared (Figure 2): Patterns 1 and 2 calcification was for the intake of sodium phosphate pur- (clumpy and finely granular) were significantly (P ¼ gatives, we performed a blinded test with a subgroup of 12 0.0133 and P ¼ 0.0007) more frequent in the calcium group cases: PAS-stained slides from four patients with phos- of diseases, whereas Patterns 3 and 4 (globular and shell- phate nephropathy after colonoscopy and eight patients like) were significantly (P ¼ 0.0032 and P ¼ 0.0014) more from the other defined classes (tumor, hyperparathyroid- frequent in the phosphate group of conditions. The differ- ism, paracellin 1 mutation, nephrotic syndrome and phos- ences of Pattern 5 (onion skin-like) were not statistically phate diabetes). The blinded set was shown to two significant, this may have been due to its very low overall experienced nephropathologists (H.H. and M.M.), who occurrence rate. Regarding these results, Patterns 1 and 2 were asked to identify the cases of phosphate nephropathy were combined and defined as the ‘calcium type’ (1 1 2) after colonoscopy by using two given criteria: (i) predom- and Patterns 3 and 4 as the ‘phosphate type’ (3 1 4) of inance of cortical calcifications and (ii) predominance of 6 T. Wiech et al. spheroidal or globular intratubular or interstitial clear de- Malignant tumours were identified in only 2 of 48 cases posits with shell-like calcifications. In the blinded test for (4.2%) in our biopsy study. In contrast, in the autopsy the identification of colonoscopy cases, one of four cases study, in 111 (49.8%) cases advanced malignant tumours was not classified as a colonoscopy case by one of the two were found. This discrepancy can be explained by the fact pathologists and one case with nephrotic syndrome was that most of the tumour patients do not undergo renal bi- wrongly classified as a colonoscopy case by both patholo- opsy because they do not suffer from renal insufficiency or gists. The overall sensitivity according to the question renal insufficiency is not the leading problem that has to be ‘which cases belong to the colonoscopy group?’ was clarified by renal biopsy. The most likely reasons for 87.5% and the specificity was 77.8%. If the question had malignancy-associated nephrocalcinosis are the presence been ‘which cases belong to the phosphate group?’ the of bone metastases (osteolytic type) or, probably less sensitivity would have been 90% and the specificity even frequently, production of PTH related peptides (humoral 100%, because the nephrotic syndrome was also classified type) by the tumour cells [13]. as a hyperphosphaturic condition in our study. The main result of our study is the observation that two In regard to all cases, including those with unknown different major types of renal calcification can be distin- etiology, we found that 24 cases (50%) of all 48 biopsies guished simply by light microscopy. The two patterns that with nephrocalcinosis had a predominant calcium type of characterize the calcium type show finely granular or calcifications. Twenty-one cases (44%) showed a predom- clumpy deposits. This type is associated with malignant

inance of the phosphate type, although in most of the cases tumours, hyperparathyroidism, sarcoidosis, Dent’s disease Downloaded from a likely underlying condition leading to nephrocalcinosis or paracellin-1 mutations. The phosphate type, in contrast, could not be determined retrospectively. In three cases is defined by the presence of shell-like or globular deposits (6%), we could not determine the calcification type, since and probably consists of a phosphate rich core, which may half of the foci were of the calcium and the other half of the be associated with several other molecules, surrounded by a phosphate type, respectively (Figure 1B). thin layer of calcium phosphate, Tamm-Horsfall protein, In order to further characterize the shell-like calcifica- and perhaps other substances. However, energy-filtered http://ndt.oxfordjournals.org/ tions, additional techniques were applied. Von Kossa and electron microscopy could only reveal the presence of cal- Alizarin stain demonstrated the presence of phosphate and cium, phosphor and oxygen in the shells but not within the calcium within the shell-like calcifications, respectively. optically empty centres. Thus, we cannot identify the Immunohistochemically, Tamm-Horsfall protein could be chemical composition of the cores of these calcifications shown in the shells of the deposits (Figure 3). Energy- and we cannot exclude washing out of the material in the filtered electron microscopy revealed calcium, phosphor centers during the fixation process. This phosphate type in and oxygen molecules as main components within the our study is associated with intake of sodium phosphate

shells of the phosphate-type calcifications (Figure 4). In solutions, nephrotic syndrome in children or phosphate at Universitaetsspital on October 27, 2011 contrast to our assumption that the optically empty centres diabetes (Table 2). of these calcifications might predominantly consist of phos- The two different patterns likely reflect different mech- phate, we could not detect phosphor and oxygen within the anisms of crystal formation; the calcium type could be centers but only in the shells. composed predominantly of calcium hydroxyl apatites No clear differences of the localization or sizes of cal- [14]. In contrast, the shell-like formations of the phosphate cification foci were seen between both major groups. In type seem to be quite similar to a self-organizing phenom- cases of acute phosphate nephropathy after colonoscopy, enon forming concentric-lamellated structures called however, most of the calcifications were intratubular and Liesegang rings. This similarity is especially apparent in located in the renal cortex. Apart from the different calci- the electron microscopic appearance of the deposits. The fication patterns of hypercalciuria and hyperphosphaturia, process of Liesegang phenomenon is thought to involve cases of sarcoidosis could be distinguished from other diffusion, nucleation, flocculation or precipitation and subtypes of calcium type nephrocalcinosis by the presence supersaturation. Liesegang-like rings have also been de- of epithelioid granulomatous reactions, adjacent to the scribed in cysts or in fibrotic, inflamed or necrotic tissue calcifications. of kidney, synovium, conjunctiva and eyelid [15, 16]. It is well known that nephrocalcinosis also occurs in renal allografts. Likely reasons are hyperparathyroidism Discussion [17], drug toxicity, as described for calcineurin inhibitors [18, 19] and hyperphosphaturia due to denervation [20]. The overall frequency of nephrocalcinosis, defined as renal Our observation can also be applied to nephrocalcinosis deposition of calcium-containing molecules (excluding cal- in renal allograft biopsies to distinguish hyperphosphaturic cium oxalate), observed in autopsies and in native renal from hypercalciuric forms. The phosphate type could be biopsies is low (autopsies 1.7%, biopsies 0.4%) but has associated with calcineurin inhibitors since for example, increased in the last 5 years. One reason could be an in- FK506 (tacrolimus) can lead to increased excretion of creased number of patients receiving sodium phosphate phosphate [21], whereas the calcium type could rather be solutions for bowel cleansing before colonoscopy [12]. associated with hyperparathyroidism (Table 2). However, However, this seems not to be the only cause since not whether this is really true and helpful to determine the only the phosphate type of nephrocalcinosis, which is typ- likely cause has to be elucidated in further studies. Early ical for phosphate nephropathy, but the calcium type as nephrocalcinosis in the renal allograft is sometimes asso- well has been observed more frequently in the last 5 years. ciated with hyperparathyroidism and is correlated with Histopathological patterns of nephrocalcinosis 7 Downloaded from http://ndt.oxfordjournals.org/ at Universitaetsspital on October 27, 2011

Fig. 3. Characteristics of the shell-like calcifications in the phosphate type of nephrocalcinosis after intake of sodium phosphate bowel purgatives before colonoscopy: (A) Shell-like calcifications (H&E, 3400, scale bar 20 lm). (B) Immunohistochemical staining for Tamm-Horsfall protein shows positivity in the shells of the calcification foci (3200, scale bar 50 lm). (C) Von Kossa reaction visualizes deposition of anions (phosphate) (3200, scale bar 50 lm) and (D) Alizarin reaction shows calcium depositions within the calcification foci (3100, scale bar 100 lm). (E) Electron microscopy reveals a lamellated structure of the calcifications with a very thin outer electron dense layer and some lighter concentric inner layers (31400, scale bar 10 lm). inferior outcome compared to renal allografts without Calcifications in patients who received sodium phos- calcifications [22]. However, in paediatric patients, renal phate solutions seemed to be localized predominantly in allograft calcification seems to be neither correlated with the renal cortex. In some patients with other causes of hypercalciuria nor with a negative influence on graft nephrocalcinosis, including calcium types, there was, how- function [23]. ever, also predominant cortical calcification and, moreover, 8 T. Wiech et al. Downloaded from http://ndt.oxfordjournals.org/ at Universitaetsspital on October 27, 2011

Fig. 4. Energy-filtered electron microscopy of a case with shell-like and globular calcifications of the phosphate type demonstrates the presence of (A) calcium, (B) phosphor and (C) oxygen molecules within the shells of calcifications with the respective energy loss spectrum on the right.

Table 2. Proposed classification of nephrocalcinosis into calcium and phosphate typesa

Calcium type (associated with hypercalciuria) Phosphate type (associated with hyperphosphaturia)

With hyperphosphataemia With hyperphosphataemia Tumor (osteolytic or humoral type) Sodium phosphate purgatives Hyperparathyreoidism Children with nephrotic syndrome Sarcoidosis and other granulomatous diseases Vitamin D excess Infantile hypercalcaemia Without hypercalcaemia Without hyperphosphataemia Idiopathic hypercalciuria Phosphate diabetes Dent’s disease Renal allografts Renal allografts Hyperparathyreoidism Calcineurin inhibitor toxicity aConditions and diseases likely associated with hypercalciuria (with or without hypercalcemia) and hyperphosphaturia (with or without hyperphospha- temia) classified according to the expected type of calcification (calcium type with finely granular or clumpy calcifications and phosphate type with globular or shell-like calcifications). Histopathological patterns of nephrocalcinosis 9 Downloaded from http://ndt.oxfordjournals.org/ Fig. 5. Hypothetical mechanism of migration of larger intratubular calcifications illustrated schematically (first row) and with examples of calcium type (second row) and phosphate type (third row) calcifications (PAS stain, 3400): Initially, the intratubular deposit destroys tubular epithelial cells (A) and leads to ‘defects’ of the tubular basement membrane (B). Regenerated tubular epithelial cells cover the calcification and the deposit migrates (C) out of the tubule until it is finally located in the interstitial space. Mostly, the tubular basement membrane shows no thickening or multilayering. some biopsies did not contain tissue of the . Acknowledgements. We thank Ursula Du¨rmu¨ller and Claudia Lautenschl- Thus, the distribution of the calcifications concerning cortex ager for excellent technical assistance, Prof. Dr Stephen Batsford for the critical review of the manuscript and all nephrologists, who provided or medulla in our study did not provide much information information about the patients. at Universitaetsspital on October 27, 2011 about the possible cause. Many of the calcifications seen in phosphate nephrop- Conflict of interest statement. None declared. athy were localized within the tubular lumens. In contrast, some of the calcium types of nephrocalcinosis had pre- dominant interstitial calcifications. In our opinion, this References could rather reflect a time course, in which the initial phase shows predominant intratubular calcifications and the later 1. Zollinger HU. Die Nephrocalcinose und die Kalknephrose. In: stages have predominant interstitial calcifications. This Zollinger HU ed Niere und ableitende Harnwege. Berlin, Germany: probably reflects a mechanism whereby larger calcium de- Springer-Verlag; 1966 posits destroy individual tubular cells and then move 2. Verkoelen CF, Verhulst A. Proposed mechanisms in renal tubular crystal retention. Kidney Int 2007; 72: 13–18 through the basement membranes into the interstitial 3. Lieske JC, Deganello S. Nucleation, adhesion, and internalization of space, which subsequently could cause interstitial fibrosis calcium-containing urinary crystals by renal cells. J Am Soc Nephrol (Figure 5). Since we do not know when precipitation of 1999; 10 (Suppl 14): 422–429 calcium phosphate started in our cases, we cannot prove 4. Vervaet BA, D’Haese PC, De Broe ME et al. Crystalluric and tubular the hypothesis in this study. However, the phenomenon of epithelial parameters during the onset of intratubular nephrocalcino- migration of the calcifications from the tubular lumens sis: illustration of the ’fixed particle’ theory in vivo. Nephrol Dial after epithelial overgrowth into the interstitium or the Transplant 2009; 24: 3659–3668 5. Vervaet BA, Verhulst A, De Broe ME et al. The tubular epithelium in so-called ‘exotubulosis’ was recently reported in animal the initiation and course of intratubular nephrocalcinosis. Urol Res experiments [24]. They also depicted the morphological 2010; 38: 249–256 phenomena outlined in Figure 5 in human renal biopsies. 6. Martin RR, Lisehora GR, Braxton M Jr et al. Fatal poisoning from Furthermore, smaller intratubular calcifications, which do sodium phosphate . Case report and experimental study. JAMA not affect the tubular epithelium, are most likely voided in 1987; 257: 2190–2192 the urine. 7. Markowitz GS, Nasr SH, Klein P et al. Renal failure due to acute In conclusion, simple light microscopy can distinguish a nephrocalcinosis following oral sodium phosphate bowel cleansing. calcium from a phosphate type of nephrocalcinosis, each Hum Pathol 2004; 35: 675–684 8. Desmeules S, Bergeron MJ, Isenring P. Acute phosphate nephropathy being associated with different diseases or conditions lead- and renal failure. N Engl J Med 2003; 349: 1006–1007 ing to calcification. Thus, histomorphology of nephrocalci- 9. Feinstein S, Becker-Cohen R, Rinat C et al. is nosis can help the clinician to assess the likely cause of prevalent among children with nephrotic syndrome and normal renal renal calcification. function. Pediatr Nephrol 2006; 21: 1406–1412 10 T. Wiech et al. 10. Schroeder JA, Weingart C, Coras B et al. Ultrastructural evidence of 18. Marbet UA, Graf U. Renale Nebenwirkungen der Therapie mit Cyclo- dermal gadolinium deposits in a patient with nephrogenic systemic fibrosis sporin A bei chronischer Polyarthritis und nach Knochemarkstrans- and end-stage renal disease. Clin J Am Soc Nephrol 2008; 3: 968–975 plantation. Schweiz Med Wochenschr 1980; 110: 2017 11. Keast VJ, Bosman M. New developments in electron energy loss 19. Mihatsch MJ, Gudat F, Ryffel B, Thiel G. Cyclosporine nephropathy. spectroscopy. Microsc Res Tech 2007; 70: 211–219 In: Tisher TC, Brenner BM (eds) Renal Pathology. Philadelphia, PA: 12. Markowitz GS, Stokes MB, Radhakrishnan J et al. Acute phosphate Lippincott; 1994, pp. 1641–1681 nephropathy following oral sodium phosphate bowel purgative: an 20. Straub B, Muller M, Heicappell R et al. Hyperphosphaturia underrecognized cause of chronic renal failure. J Am Soc Nephrol after kidney transplantation in syngeneic rats: effects on nephro- 2005; 16: 3389–3396 calcinosis and bone metabolism? Transplant Proc 2003; 35: 13. Weiss M, Liapis H, Tomaszewski JE et al. and other 1575–1580 infections, reflux nephropathy, , and nephrolithiasis. 21. Mohebbi N, Mihailova M, Wagner CA. The calcineurin inhibitor In: Jennette JC, Olson JL, Schwartz MM, Silva FG (eds) Heptinstall’s FK506 (tacrolimus) is associated with transient metabolic acidosis Pathology of the Kidney. Philadelphia, PA: Lippincott Williams & and altered expression of renal acid-base transport proteins. Am J Wilkins; 2006 Physiol Renal Physiol 2009; 297: 499–509 14. Zollinger HU, Mihatsch MJ. Nephrocalcinosis. In: Renal Pathology in 22. Gwinner W, Suppa S, Mengel M et al. Early calcification of renal Biopsy. Berlin, Germany: Springer Verlag, 1978, pp. 483–486 allografts detected by protocol biopsies: causes and clinical implica- 15. Tuur SM, Nelson AM, Gibson DW et al. Liesegang rings in tissue. tions. Am J Transplant 2005; 5: 1934–1941 How to distinguish Liesegang rings from the giant kidney worm, 23. Habbig S, Beck BB, Feldkotter M et al. Renal allograft calcification— Dioctophyma renale. Am J Surg Pathol 1987; 11: 598–605 prevalence and etiology in pediatric patients. Am J Nephrol 2009; 30: 16. Sneige N, Dekmezian R, Zaatari GS. Liesegang-like rings in fine 194–200 needle aspirates of renal/perirenal hemorrhagic cysts. Acta Cytol 24. Vervaet BA, Verhulst A, Dauwe SE et al. An active renal crystal Downloaded from 1988; 32: 547–551 clearance mechanism in rat and man. Kidney Int 2009; 75: 41–51 17. Schwarz A, Mengel M, Gwinner W et al. Risk factors for chronic allograft nephropathy after renal transplantation: a protocol biopsy study. Kidney Int 2005; 67: 341–348 Received for publication: 30.11.10; Accepted in revised form: 17.6.11 http://ndt.oxfordjournals.org/ at Universitaetsspital on October 27, 2011